Application Of Time Series Analysis To Measurement Of Displacements In A Tunnel

As an important embranchment of probability and statistics, time series analysis has been regarded as a powerful and potential tool for prediction and control. It has been developed rapidly in recent years, especially in physical science, social science and engineering technology. To fix ranks is one of the key steps in the analyses based on time series.In this thesis, the conventional methods commonly used to fix ranks for the time series are examined. The merits and drawbacks of these methods are evaluated. Further investigations on the AR model of the time series are made on the basis of the measured data of displacements obtained from a construction of tunnel by NATM and the T-F method is developed for fixing ranks. It is proven through an example analysis to be suitable and available for the NATM construction of tunnel. Taking the tunnel displacement of the Zhongliang mountain of Sichan province as an example, an AR(7) model is established and then employed for predictions of deformations. Compared with the original observations, the prediction is of high accuracy. Accordingly, the method proposed in this thesis is shown to be widely applicable to engineering practice.As the second part of the thesis, considering the fact that the accuracy of the measurement of fractal dimensions are usually insufficient, a procedure for modifying the trial data based on probability analysis and statistics is proposed in order to reduce trial errors and improve the precision of fractal dimensions. In addition, by practical proof-tests of trial measurements, it is manifested that the methodology is basically feasible.